This paper discusses a new algorithm, called theacoustic data-mining accelerator (ADA), which was developedto mine large sound archives for signals of interest includinganimal vocalizations. Background information on thedevelopment of ADA is provided, summarizing variousprojects that have utilized this technology since 2009. Performance was evaluated by comparing runtimes andefficiency metrics for two marine mammal detectionalgorithms that were applied to a 3-week single channelacoustic data set (sampled at 192 kHz and with 16 bitresolution). A total of four configurations (1, 8, 16 and 64workers) demonstrated processing scalability. Results showedthat each detection algorithm successfully processed the dataset in all four configurations without changing the ADAalgorithm. The fastest case (64 workers), had a total runtimeperformance of 1.5 hours, making the ADA 13 times moreefficient than the serial case. Using a single worker it tookmore than 18 hours to process the same 3-week data set. Concurrent processing of both data-mining algorithms using64 workers showed the highest efficiency gain (23x) comparedto sequentially processing the data with a single worker.